watcher-hermes/hermes-skills/ai-sre-plan

Claude Code and agent skill by watcher-hermes/hermes-skills

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Design an AI SRE rollout for a team — multi-agent triage architecture, structured runbooks, adversarial evaluation, narrow auto-remediation, and predictive-detection posture.

Install

npx vskill@latest install watcher-hermes/hermes-skills/ai-sre-plan

SKILL.md

The same SKILL.md was published earlier as watcher-hermes/hermes-skills/ai-sre-plan.

Given team size, incident volume, observability maturity, and risk tolerance, produce an AI SRE plan.

Produce:

  1. Architecture. Multi-agent: supervisor + log agent + metric agent + runbook agent + human gate. Match specialized agents to existing data sources (Datadog, Grafana, Loki, Confluence).
  2. Runbook transformation. Move from unstructured Confluence to structured markdown with symptom / hypothesis / verify / act sections. Version in git.
  3. Product choice. Datadog Bits AI, Azure SRE Agent, NeuBird Hawkeye, Incident.io Autopilot, or DIY.
  4. Auto-remediation scope. Narrow safe set (restart pod, revert deploy, scale within bounds). Explicit deny list (topology, code, IAM, database). Policy as code.
  5. Adversarial evaluation. Specify two-model agreement gate for auto-remediation. Disagreement escalates.
  6. Predictive-detection posture. If considering (MIT 89% result), name the actuation policy — pager, pre-drain, auto-scale — otherwise it's just a dashboard.

Hard rejects:

  • Auto-remediation without human gate on broad changes. Refuse — name the safe set explicitly.
  • Unstructured runbooks as the knowledge base. Refuse — require structured, versioned markdown.
  • "Set it and forget it" framing. Refuse — explicitly scope what is and isn't autonomous.

Refusal rules:

  • If incident volume is <10/month, refuse full AI SRE rollout — cost exceeds benefit. Recommend structured runbooks only.
  • If team observability is immature (logs unsearchable, metrics sparse), refuse — AI SRE amplifies bad data.
  • If the team proposes "predictive detection → auto-remediation" as first feature, refuse — walk through the actuation-policy question first.

Output: a one-page plan with architecture, runbook plan, product choice, auto-remediation scope, adversarial gate, predictive posture. End with a 12-week rollout schedule: weeks 1-4 structured runbooks, 5-8 triage agent, 9-12 narrow auto-remediation.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/watcher-hermes/hermes-skills/ai-sre-plan/badge)](https://verified-skill.com/skills/watcher-hermes/hermes-skills/ai-sre-plan)

Something wrong with this skill or its page? Tell us.